Human–Machine Interactions would significantly benefit from systems capable of predicting human emotions, potentially revolutionizing user experience and outcomes. The ‘PHANTOMATRIX’ project explores the use of Machine Learning (ML) to predict emotional states by analyzing biological signals such as heart rate and facial expressions within Virtual Reality (VR) environments. This study utilizes VR’s immersive capabilities for evoking and measuring emotions through physiological, kinematic, and self-reported data, employing wearables and cameras. Although previous research has highlighted ML’s potential in predicting physiological responses to VR stimuli, the precise indicators of emotion remain elusive to identify due to the subjective nature of emotions and overlapping physiological signs. ‘PHANTOMATRIX’ addresses this by measuring individual emotions across multiple modalities linked to specific VR scenarios, significantly extending current research. The project focuses on the effectiveness of VR environments in eliciting focus and relaxation, correlating physiological responses to VR parameters, and using these parameters in ML models to predict emotional states. The outcomes aim to significantly enhance our understanding of mechanisms for emotion induction, classification, and regulation, with broad implications in societal, educational, and health sectors, particularly improving Human–Machine Interaction.